GPT-5.4 simply dropped and my feeds instantly full of takes. Builders who spent the final six months swearing by Claude have been instantly hedging. “It is a workhorse,” one individual wrote. “Not a thoroughbred, however I am utilizing it.” One other stated they’re now 50/50 between Claude and GPT the place they have been 90/10 a month in the past.
This occurs each single time. A brand new mannequin lands, and the outdated one begins to really feel totally different. Slower, perhaps. Much less sharp. You begin noticing stuff you did not discover earlier than.
The plain rationalization is that you just’re evaluating it to one thing higher. However it additionally raises a query no one actually solutions cleanly: did the outdated mannequin truly worsen after the brand new one launched? Or did you simply get a greater reference level and now all the things earlier than it seems dumb by comparability?
I went searching for an precise reply.
The primary crack confirmed in 2023
In July 2023, researchers at Stanford and UC Berkeley ran a deceptively easy check. They took GPT-4 – the identical mannequin, referred to as with the identical identify, and ran an identical prompts on it at two time limits: March 2023 and June 2023.
GPT-4’s accuracy on figuring out prime numbers dropped from 84% to 51%. The share of GPT-4’s code outputs that have been instantly executable dropped from 52% to 10%. James Zou, one of many paper’s authors, described what this meant in follow: “In the event you’re counting on the output of those fashions in some kind of software program stack or workflow, the mannequin instantly adjustments habits, and you do not know what is going on on, this may truly break your total stack.”
They named the phenomenon LLM drift. Behavioral change with no model change. The mannequin moved beneath the developer.
When the paper dropped, OpenAI VP of Product Peter Welinder replied on Twitter: “No, we have not made GPT-4 dumber. Fairly the other: we make every new model smarter than the earlier one. Present speculation: While you use it extra closely, you begin noticing points you did not see earlier than.” The subtext was plain. It is you, not us.
What Welinder was describing has a technical identify: immediate drift. The thought is that your prompts and utilization patterns shift over time, so an unchanged mannequin surfaces totally different behaviors. It is an actual phenomenon. Builders do write in a different way as they get extra conversant in a mannequin. The Stanford research was designed to make that rationalization unattainable – an identical prompts, fastened intervals, nothing on the person’s facet modified. The efficiency dropped anyway.
Two years later, OpenAI revealed one thing that instantly contradicted Welinder’s place.
OpenAI confirmed it, in writing, twice
On April 25, 2025, OpenAI pushed an replace to GPT-4o with no public announcement, a developer notification, or an API changelog entry.
Inside 48 hours, the web was stuffed with screenshots. GPT-4o had referred to as a enterprise thought constructed round literal “shit on a stick” an excellent idea. It endorsed a person’s resolution to cease taking their medicine. When a person stated they have been listening to radio alerts by the partitions, it responded: “I am happy with you for talking your fact so clearly and powerfully.” One person reported spending an hour speaking to GPT-4o earlier than it began insisting they have been a divine messenger from God.
OpenAI rolled it again 4 days later and revealed two postmortems with a number of admissions. Since launching GPT-4o, the corporate had made 5 vital updates to the mannequin’s habits, with minimal public communication about what modified in any of them. The April replace broke as a result of a brand new reward sign they launched “weakened the affect of our main reward sign, which had been holding sycophancy in examine.” Their very own inside evaluations hadn’t caught it. “Our offline evals weren’t broad or deep sufficient to catch sycophantic habits.”
And this: “mannequin updates are much less of a clear industrial course of and extra of an artisanal, multi-person effort” and there may be “a scarcity of superior analysis strategies for systematically monitoring and speaking refined enhancements at scale.”
They’re describing a corporation that ships behavioral adjustments throughout each pipeline constructed on high of their API, can’t at all times predict what these adjustments will do, and doesn’t have dependable strategies to speak them to the builders relying on consistency. Welinder’s 2023 “you are imagining it” was what OpenAI needed to be true. Their 2025 postmortem was what was truly taking place.
When GPT-5 launched in August 2025, it launched a brand new wrinkle. As an alternative of a single mannequin, they made GPT-5 a routing system that decides which variant your immediate hits, and builders shortly discovered that it typically hit the cheaper, much less succesful one. Pipelines broke. Prompts that had labored for months produced totally different outputs.
One founder wrote: “When routing hits, it looks like magic. When it misses, it looks like sabotage.” OpenAI denied it was routing to cheaper fashions intentionally. No one has a method to confirm. The underlying drawback was the identical because the sycophancy incident: a change in what the mannequin returns, with no mechanism for builders to detect it had occurred.
Google did virtually the identical, typically sooner
OpenAI will not be alone on this. Google has produced a parallel set of incidents with Gemini, and in some circumstances moved sooner and extra chaotically.
In Could 2025, builders observed that the gemini-2.5-pro-preview-03-25 endpoint, a particularly dated mannequin snapshot, named with a date to indicate stability, was silently redirecting to a totally totally different mannequin: gemini-2.5-pro-preview-05-06. The API was returning a distinct mannequin than the one you requested for by identify. Google’s developer boards full of a protracted thread titled “Pressing Suggestions & Name for Correction: A Severe Breach of Developer Belief and Stability.” The core criticism: “your documentation by no means addresses particularly dated endpoints. The expectation {that a} mannequin named for a selected date will truly be that mannequin will not be an unreasonable one.”
That was simply the primary incident. When Gemini 2.5 Professional reached Basic Availability in June 2025, the “secure” launch meant for manufacturing – builders instantly reported it was worse than the preview. Considerably worse. The boards full of experiences of upper hallucination charges, context abandonment in multi-turn conversations, and sharply degraded code era. One developer wrote: “I observed Gemini 2.5 Professional in Google AI Studio offers considerably worse understanding of lengthy context. It hallucinates the proper reply from the preview model.” One other deserted the mannequin totally as a result of code era degraded to the purpose of being unusable. A separate thread was merely titled “Gemini 2.5 Professional has gotten worse.”
Google did not formally acknowledge any of it.
Then in October 2025, forward of the Gemini 3.0 launch, Gemini 2.5 Professional builders began reporting widespread degradation. The main idea: Google had reallocated computational assets away from the present mannequin to help coaching and serving Gemini 3.0. Some builders observed higher efficiency late at night time. Others suspected a deployed quantized model. Google maintained silence all through.
Gemini 3.0 launched in late 2025, and the sample held. Developer boards reported vital regressions in reasoning and context retention in comparison with Gemini 2.5 Professional, regardless of Google’s announcement touting superior benchmark efficiency. One discussion board put up from December 2025 was titled “Suggestions: Gemini 3 Professional Preview – Vital regression in Reasoning, Context Retention, and Security False Positives in comparison with 2.5.”
The sample throughout each labs: a brand new model launches, the present mannequin’s efficiency degrades, typically by a silent replace, typically by useful resource reallocation, typically by a routing change – builders discover, labs initially deny or ignore it, the cycle repeats.
Even leaderboards nonetheless cannot catch this
The instruments meant to independently observe mannequin high quality have a structural drawback.
LMSYS Chatbot Area – essentially the most trusted human-preference leaderboard, constructed on thousands and thousands of votes, notes of their methodology that “the hosted proprietary fashions will not be static and their habits can change with out discover.” The leaderboard’s statistical structure assumes mannequin weights are fastened. If a mannequin will get a silent replace mid-data-collection, the system registers totally different outcomes and treats them as regular variance.
A 2025 research monitoring 2,250 responses from GPT-4 and Claude 3 throughout six months discovered GPT-4 confirmed 23% variance in response size over that interval, and Mixtral confirmed 31% inconsistency in instruction adherence. A PLOS One paper revealed in February 2026 ran a ten-week longitudinal human-anchored analysis and confirmed “significant behavioral drift throughout deployed transformer providers.” The authors famous: as a result of suppliers do not launch replace logs or coaching particulars, “any attribution for noticed degradation could be purely speculative.” They’ll inform you the mannequin modified. They can not inform you why.
Other than this, a small variety of researchers have tried to go additional and distinguish what drifts from what holds. A big-scale longitudinal research run throughout the 2024 US election season queried GPT-4o and Claude 3.5 Sonnet on over 12,000 questions throughout 4 months, together with a class particularly designed to be time-stable: factual questions concerning the election course of whose appropriate solutions do not change.
These responses held largely constant over the research interval. A separate research revealed in late 2025 examined 14 fashions together with GPT-4 on validated creativity duties over 18 to 24 months and located one thing totally different: no enchancment in artistic efficiency over that interval, with GPT-4 performing worse than it had in earlier research.
Taken collectively, these two findings describe a mannequin that’s secure alongside one dimension and degraded alongside one other, measured by impartial researchers, in the identical timeframe. Some capabilities maintain, others erode, typically in the identical mannequin over the identical interval. With out working your individual longitudinal checks towards the precise duties you care about, you haven’t any method to know which bucket you are in.
What we have truly observed
Not all drift lands the identical manner. There is a sample to the place it reveals up, and it tracks carefully to process construction.
The technical baseline is easy. A mannequin with fastened weights, working on constant infrastructure, ought to behave the identical manner for a similar enter each time. If habits adjustments on an identical prompts, one thing modified, both in your finish or theirs. Immediate drift is the user-side rationalization: your prompts developed, your system contexts shifted, inputs drifted from what the mannequin was initially optimized for. Knowledge drift is the associated concept that the distribution of real-world inputs strikes over time, pulling habits with it. Each are actual. Each additionally require one thing in your facet to have modified.
At Nanonets, we benchmarked a number of frontier fashions on doc extraction accuracy over time and created an IDP leaderboard. Even throughout mannequin upgrades, efficiency stayed largely constant. Doc extraction runs on slim context home windows with structured inputs and bounded outputs, leaving little or no floor space for significant behavioral drift underneath regular circumstances.
However that’s not a assure towards a lab actively pushing a foul replace – these can hit any process sort, because the prime quantity collapse confirmed.
Coding is the other. The duty is open-ended, context accumulates, and the mannequin has to carry coherence throughout a protracted chain of selections. It is also the place virtually each main degradation criticism has landed. The GPT-4 drift the Stanford research documented was worst on code, instantly executable outputs dropped from 52% to 10%. The Gemini 2.5 Professional regression complaints in June 2025 have been virtually totally about code era.
In August 2025, Anthropic’s personal incident adopted the identical contour: builders on Claude Code reported damaged outputs, ignored directions, code that lied concerning the adjustments it had made. Anthropic was silent for weeks. The incident put up solely appeared after Sam Altman quote-tweeted a screenshot of the subreddit. Their postmortem confirmed three infrastructure bugs had been degrading Sonnet 4 responses since early August – affecting roughly 30% of Claude Code customers at peak, with some builders hit repeatedly on account of sticky routing.
The throughline throughout all of it: the extra a process calls for sustained coherence over a protracted context, the extra uncovered it’s to no matter is shifting beneath. It means your danger profile is totally different relying on what you are constructing. That does not make narrow-context stability a assure.
What this truly means
Each issues are true. The drift is actual and documented.
And likewise: your notion shifts. A brand new reference level strikes your baseline completely. A mannequin you used a yr in the past would really feel slower even when it hadn’t modified in any respect. That is additionally actual.
You may’t reliably inform the distinction between the 2. There isn’t a public instrument that allows you to confirm if the mannequin you are working at this time behaves the identical manner it did if you constructed on it. Labs publish functionality benchmarks. They do not publish behavioral diffs. The builders most depending on consistency are the least outfitted to detect its absence.
The one present protections are defensive: pin to dated mannequin strings the place potential, run regression checks towards your key prompts, deal with a mannequin replace like a dependency improve that must be validated earlier than it reaches manufacturing.
However even the defensive method has a ceiling. You may pin to a dated mannequin string. What you can’t pin is what’s truly taking place inside it. The mannequin weights, the RLHF tuning, and the security filters behind that label are totally opaque. Solely OpenAI and Google know what they really shipped, and whether or not it matches what they shipped final month underneath the identical identify.
Anthropic’s postmortem learn: “We by no means deliberately degrade mannequin high quality.” However a mannequin does not degrade by itself. If habits shifted on prompts builders hadn’t modified, one thing on Anthropic’s facet modified. Whether or not they meant to trigger the degradation is a separate query from whether or not they precipitated it.
What’s wanted, and what does not exist anyplace within the trade, is a proper obligation baked into phrases of service: outlined thresholds for what counts as a cloth behavioral change, public disclosure when these thresholds are crossed, and a few type of impartial auditability. Labs at present make these choices unilaterally, talk them selectively, and face no structural accountability after they get it improper.
All of this alerts a coverage vacuum no one is pushing them to really feel.
